This study will test the hypothesis that opioids contribute to postoperative sleep disturbance. Additionally, this study will determine whether the adverse effects of sleep disturbance on the cardiovascular, respiratory and central nervous systems are less pronounced in patients treated with regional local anesthetic instead of opioids for postoperative pain control. Finally the hormone condition corresponding to sleep disruption will be examined to evaluate the influence of stress hormones and the response of melatonin, a hormone controlled by the circadian rhythym. The primary aim is to determine if opioids contribute to postoperative sleep disturbance. Patients undergoing abdominal hysterectomy will be randomly assigned to receive epidural opioid or local anesthetic for postoperative pain control. The patients' sleep will be monitored on the night before surgery and on the first three posoperative nights by recording the electroencephalogram and electromyogram. For each patient the amount of REM sleep and NREM sleep on the preoperative night will be compared with the amounts on the postoperative nights. We expect to observe greater differences before and after surgery in theses sleep stages in the opioid group, We expect preservation of a more normal sleep cycle in the local anesthetic group. Because surgical stress and pain are potential contributers to sleep disruption, we will measure stress hormones in the urine (catecholamines and cortisol) and blood (interleukin-1) as well as record the patients' subjective rating of their pain. The second aim is to determine if postoperative sleep disturbance is clinically significant. Studies have shown that sleep deprivation with consequent REM rebound can cause impaired cognitive performance, breathing disturbance, and hemodynamic instability. Daily tests of cognitive function and overnight recordings of respiratory rate, arterial oxygen saturation, blood pressure and the electro- cardiogram will be performed. We hypothesize that the complications of sleep deprivation demonstrated in normal volunteers will occur in postoperative patients who are already under intense physiologic stress. The final aim is to test the hypothesis that melatonin, a hormone controlled by the circadian rhythym, will be supressed during the nights of sleep deprivation. Abnormally low levels of melatonin have been measured in elderly insomniacs, and administration of melatonin to these patients has improved their sleep. If postoperative sleep distruption is associated with abnormally low levels of melatonin in the blood, then administration of melatonin might be a potential treatment for postoperative sleep disturbance.